"Statistical Significance" is a term that is used in almost all quantitative studies that use inferential statistics (ANOVA, t-tests, regression). However, for practical considerations, there might be an overreliance on the "victory" of achieving statistical significance. Why? Because even though the results may be statistically significant, they may not be of practical value. For example, let's assume that the pre-test average score of a group of students is 74 and the post-test average score is 76. In fact, let's even assume that we are 95% confident, from our tests, that the two groups are different from something other than normal process variance. However, is two points of improvement on a test meaningful from a practical perspective? In other words, did the students really achieve greater mastery at a practical level? I think that most practitioners would agree with me that the answer is: "no." This is when the effect size analysis is of value for the scholar-practitioner.

There is much to discuss on this topic, but I will leave technical explanation to future blogs - the purpose of this blog is to encourage you to start to think about the concept of effect size. In upcoming blogs, I will be writing more about power analysis, effect sizes, and statistical significance. I believe that the PowerPoint at the link below is an excellent follow-up resource for learners who want to learn more about this topic.

http://www.slideshare.net/jtneill/power-effect-sizes-confidence-intervals-scientific-integrity

Daniel J. Smith, MA, MBA, PhD

480.861.8851 (cell)

[email protected]